On Wed, 10 Feb 2016 17:56:00 +0100, Martin Natano wrote: > What do you mean by it being used in a clever way? I see the sign bit > being used as part of the hash itself, like all the other bits in the > sum. That shouldn't change when using unsigned instead. Am I missing > something?
I was trying to go a bit further and fix other sign compare issues and fell down the rabbit hole. See the comment about equiv at the top of the file. > For what it's worth, converting the hash to unsigned at least doesn't > seem to break anything obvious for me. See the diff below (generated > with the patched binary). Just changing the hash value to unsigned is probably OK as it is only used to compare to otherh hash values. - todd
